A Fast-Processing Predictive Control Strategy for Common-Mode Voltage Reduction in Parallel Three-Level Inverters

Model predictive control (MPC) is an attractive control scheme, in particular, for multilevel converters. However, MPC suffers from high computational burden and tedious tuning of weighting factors especially for parallel three-level inverters. Moreover, the parallel three-level inverters will unavoidably face the challenge of maintaining low circulating current, reducing common mode voltage (CMV), and balancing neutral point (NP) voltage in practice. Therefore, in this article, a fast-processing predictive control strategy is proposed for reducing the zero-sequence circulating current (ZSCC), NP voltage imbalance and CMV simultaneously in parallel three-level inverters. Based on the selected sector by the reference voltage vector, the proposed fast-processing predictive control takes feasible five or six candidate vectors into account; therefore, the computational burden, which is a disadvantage of the predictive control for inverters, will decrease. In addition, the proposed method eliminates the weighting factor, which avoids tedious tuning work. In the proposed control scheme, the candidate vectors are selected for ZSCC, NP voltage imbalance, and CMV reduction, and then, among them, the optimized switching vector is selected by optimizing the cost function. The performance of the proposed method has been validated by the simulation and experimental results.